Clinical protein assays: Enrichment effects on interlab reproducibility in SRM studies

Measuring proteins in clinical samples

The necessity for estimating the levels of particular proteins in clinical samples like serum or tissue is growing rapidly as more and more protein isoforms associated with certain diseases are being identified. The conventional method using immunoassays for each particular protein generally works well and can be set up on automated systems. However, in some cases these assays are not specific to the target protein and they can also suffer from poor reproducibility between different labs.

An obvious alternative is mass spectrometry, particularly the use of selected reaction monitoring (SRM). This technique has high specificity for a particular amino acid sequence within a protein, which is an important aspect to consider when protein isoforms are being considered as it can differentiate between proteins of similar structure. It is also blessed with speed and high-throughput capabilities through multiplexing.

Despite these welcome characteristics, SRM has also been criticised for its poor interlab reproducibility, suggesting that a test in one lab cannot always be reproduced in another lab. Nonetheless, there have been several recent studies which contradict this perception, especially when standard operating procedures and standard isotope-labelled peptides were employed.

One potential complication with SRM appears when low-abundance proteins are targeted. In these cases some sort of enrichment is required to concentrate the proteins so that they can be measured accurately, but how will this extra step affect the interlab reproducibility? One new project has addressed this question by performing an interlab study as described by senior reporter Amol Prakash from Biomarker Research in Mass Spectrometry, Thermo Fisher Scientific, Cambridge, MA.

Enrichment effects compared

Working with a team of coresearchers from five universities in the US, Canada and China, Prakash summarised their strategy. Three different sample enrichment procedures were used before the samples were analysed using established standard operating procedures. Each lab processed clinical samples by one of the methods and distributed samples to the other three labs for analysis

Firstly, human serum from normal and prostate cancer patients and healthy female serum spiked with prostate-specific antigen were digested with trypsin and the N-linked glycopeptides were captured on a hydrazide resin. The second method involved the isolation of prostate-specific antigen from human serum by immunoassay before tryptic digestion. Finally, pituitary human growth hormone was enriched from urine on nanoparticles, followed by tryptic digestion.

Each sample was analysed by LC/MS/MS following electrospray ionisation and specific SRM transitions were established for characteristic peptides derived from each target protein. The amounts of the peptides representing the target proteins were compared across all of the labs from each enrichment procedure.

The team found that the SRM methods were sufficiently sensitive to measure the peptides at clinically relevant levels and in a reproducible manner between the labs. In all cases, the interlab CVs were less than 30%, falling as low as 5% in some cases, figures which are well within acceptable limits. So, the mode of enrichment did not notably affect the amounts of proteins measured in different labs

One key factor in this comparison was that all of the assays were carried out on the same model of triple quadrupole mass spectrometer. The introduction of different types of mass spectrometer might have an effect on the reproducibility, so the process would need to be examined again.

Accurate measurement of proteins in clinical samples like plasma and urine will become more important in the future, especially with the emergence of personalised medicine. The excellent reproducibility of mass spectrometry in SRM mode, combined with the speed and multiplexing ability, suggests that this technique can be applied successfully, even when enrichment of the target proteins is required.